Q4: Model card updates

[AkshitaB]

Prioritized fields:

• Model details
• Metrics
• Evaluation data
• Training data

Fill in the prioritized fields in all model cards:

• [x] coref-spanbert
• [x] generation-bart
• [x] glove-sst
• [x] lm-masked-language-model
• [x] lm-next-token-lm-gpt2
• [x] mc-roberta-commonsenseqa
• [x] mc-roberta-piqa
• [x] mc-roberta-swag
• [x] pair-classification-decomposable-attention-elmo
• [x] pair-classification-esim (Missing/old information!)
• [x] pair-classification-roberta-mnli
• [x] pair-classification-roberta-snli
• [x] rc-bidaf
• [x] rc-bidaf-elmo
• [x] rc-naqanet
• [x] rc-transformer-qa
• [x] roberta-sst
• [x] structured-prediction-biaffine-parser
• [x] structured-prediction-constituency-parser
• [x] structured-prediction-srl
• [x] structured-prediction-srl-bert
• [x] tagging-elmo-crf-tagger
• [x] tagging-fine-grained-crf-tagger
• [x] tagging-fine-grained-transformer-crf-tagger

Additional models in the demo:

• [x] All semparse models
• [x] NMN

Remaining fields

• Intended use
• Factors
• Quantitative analyses (wrt chosen factors).
• Ethical considerations
• Caveats and recommendations

Misc Notes:

• Create a list of datasets that are used, and information about them.
• Should a new field similar to model_type be added, which specifically mentions the underlying pretrained model? Eg. if a model is finetuned on BERT, it is susceptible to the same limitations as BERT.

[schmmd]

Model card information is stored here: https://github.com/allenai/allennlp-models/tree/master/allennlp_models/modelcards

[github-actions[bot]]

@AkshitaB this is just a friendly ping to make sure you haven't forgotten about this issue 😜

[github-actions[bot]]

@AkshitaB this is just a friendly ping to make sure you haven't forgotten about this issue 😜

[github-actions[bot]]

@AkshitaB this is just a friendly ping to make sure you haven't forgotten about this issue 😜